Why not craft the inner packet with inner src and destination to a local loopback address (127.0.0.1) before sticking it in the UDP header? Preserves layering, preserves multihoming, avoids public/private address space concerns, prevents similar tunnelling by middleboxes. Of course, there's tracking the real src and dst in the added/discarded UDP header, and associating those with the decapped inner packet -- tch, implementation details... L. -----Original Message----- From: Phelan, Tom [mailto:tphelan@xxxxxxxxxxxx] Sent: 27 April 2010 18:46 To: Alex Conta; Wood L Dr (Electronic Eng); fred@xxxxxxxxx Cc: dccp@xxxxxxxx; tsv-area@xxxxxxxx; tsvwg@xxxxxxxx Subject: RE: UDP encaps for SCTP and DCCP -- why not just tunnel it? Hi All, Well, quite a few issues have been brought up here. Here are my thoughts on one of them. Why not just tunnel it? "Just tunnel it" means IP header/UDP header/inner IP header/"real" transport header. The problem here is in the addresses in the inner IP header. In most uses of IPv4 tunneling, the tunnel endpoints either both have public IP addresses or they both exist in the same private IP address space. This provides a coordination of the inner addresses such that they are meaningful to both the sender and receiver. In these conditions, DCCP can be tunneled just like everything else. However, there are many situations where these conditions don't apply. One very common situation where these conditions don't exist is when the client, managed by one organization, is in a private address space behind a NAT and the server, managed by another organization, is in the public address space. This situation, and the several variants that are possible, are more the target of the DCCP encap draft. In the private-client-nat-public-server configuration, the client will initially set the outer and inner source addresses to private addresses (possibly the same address for both, possibly different). The NAT will change the outer source address to a public address, but not the inner. When the inner packet is extracted at the server, the inner source address will be meaningless (and leaks information that is normally held private). On the return path, it'll be the destination address that's meaningless. You can imagine many ways that the client can set inner addresses to values that would mean something to the server, but it's difficult to see how the server could set the inner addresses meaningfully without having the client's private address leaked to it (define a "neutral" address space?). But if you do solve the inner address meaning problem, what have you gained? Nothing, I believe. The inner addresses are fundamentally useless. So why carry useless fields? Make it IP header/UDP header/transport header. This is the approach all three current drafts take. Tom P. > -----Original Message----- > From: tsv-area-bounces@xxxxxxxx [mailto:tsv-area-bounces@xxxxxxxx] On > Behalf Of Alex Conta > Sent: Tuesday, April 27, 2010 9:54 AM > To: L.Wood@xxxxxxxxxxxx; fred@xxxxxxxxx > Cc: dccp@xxxxxxxx; tsv-area@xxxxxxxx; tsvwg@xxxxxxxx > Subject: RE: UDP encaps for SCTP and SCCP > > IMHO, what should drive the use or no use of UDP is "the need and use of > UDP functons", rather then "the need to make the SCTP and SCCP more > popular because of using a popular transport". Perhaps I am not reading > right the motivation, but I somehow get that frm the thread - sorry... > > Which is to say that I agree with Fred's concern 100% about duplications > of functions. > > I would perhaps even go a little further, if the counter-argument is > that with "UDP lite", there is a minimization of duplication, and that > is: > > my concern would be that besides the dupplication, there is some > additional burden in processing a packet that has mutliple header > layers, when passing the packet from one layer to the other. It's not > much, but there is, and can add-up - there are the little things that, > when in larhe numbers, contribute to a considerable load. > If that can be eliminated, I see it as a plus... > > Alex Conta > > > -----Original Message----- > > From: tsv-area-bounces@xxxxxxxx > > [mailto:tsv-area-bounces@xxxxxxxx] On Behalf Of L.Wood@xxxxxxxxxxxx > > Sent: Tuesday, April 27, 2010 6:38 AM > > To: fred@xxxxxxxxx > > Cc: dccp@xxxxxxxx; tsv-area@xxxxxxxx; tsvwg@xxxxxxxx > > Subject: RE: UDP encaps for SCTP and SCCP > > > > > > Fred, > > > > advocating removing the interior protocol checksum across header and > > payload (and reconstituting/recomputing it > > afterwards) violates end-to-endidness and weakens the reliability > > guarantee. Bad idea. From a checksum perspective, you'd do better > > saying 'use UDP lite with a minimal check just across its own > > headers and pseudo-header' to decrease the computational overhead - > > that should be a fixed value. > > > > Using a UDP (or lite) port to indicate what is being done here with > > a particular weird encap, as well as the original ports on the > > interior packet, is something that this approach is stuck with, I > > think. > > > > Lloyd Wood > > http://sat-net.com/L.Wood ________________________________________ > > From: Fred Baker [fred@xxxxxxxxx] > > Sent: 27 April 2010 10:55 > > To: Lars Eggert > > Cc: bidulock@xxxxxxxxxxx; dccp@xxxxxxxx; tsv-area@xxxxxxxx; tsvwg > > list; Wood L Dr (Electronic Eng); Michael Welzl > > Subject: Re: UDP encaps for SCTP and SCCP > > > > From my perspective, I would prefer to run a native encapsulation > > rather than host it in UDP. If one wants a UDP encapsulation, I have > > no opinion on which of the choices to make, but I would suggest a > > characteristic you want to have. > > There is no point having UDP ports *and* SCTP/DCCP ports, and no > > point in having a UDP checksum *and* an SCTP checksum. I would > > recommend removing the duplicated functions from the interior > > protocol and relying on UDP's counterpart, even if it is inferior, > > as it will be more readily deployed. > > > > On Apr 27, 2010, at 10:42 AM, Michael Welzl wrote: > > > > > Hi, > > > > > > Okay, I herewith speak up: yes I want to see UDP encapsulation for > > > both these protocols (but right now I'm not sure which one). > > > > > > Both SCTP and DCCP are useful - if there was no consensus on that, > > > ever, these groups would never have been formed, and the protocols > > > would never have been developed. > > > > > > Now, they are not used much (on the Internet involving > > NATs); at least > > > DCCP isn't. That's a problem. UDP encapsulation is a way to try to > > > solve this problem - and saying that we shouldn't do this > > because the > > > protocols aren't used is a bit stupid, isn't it? > > > > > > To repeat this more clearly and bluntly: > > > > > > tool X isn't working well => noone uses it. > > > So let's not fix tool X because noone uses it anyway. > > > Hmmm... > > > > > > Cheers, > > > Michael > > > > > > > > > On Apr 27, 2010, at 9:47 AM, Lars Eggert wrote: > > > > > >> Hi, > > >> > > >> please keep this discussion focused on which approach we should > > >> follow for UDP-encapsulating DCCP and SCTP. > > >> > > >> I'm happy Lloyd posted his views. I'm hoping other > > community members > > >> will speak up as well. If I were asked to characterize current > > >> consensus, I'd probably say "disinterest for either > > approach." (Which > > >> would be fine, but doesn't quite match the earlier feeling > > I got from > > >> the community, i.e., that we do want UDP encaps for these > > protocols.) > > >> > > >> Lars > > > > > > http://www.ipinc.net/IPv4.GIF > >
